Try again with formatting...

Hi all,
this is our report about possible specification for "links to the rest
of the world" as recorded in [7]

General remarks:
The main question for things that are not directly related to
emotion/affect is: shall we specify them ourself or advise the use of
similar xml specifications (make or buy)? 

arguments to use existing specifications:
  a. It is not core to Emotion, so not worth us inventing something new
  b. Many specifications exist
  c. Using a format from one of our "neighbors" will help foster closer integration of our groups / standards

arguments for own specification:
  a. other specifications have different focus: more "workaround" than
  real fit.
  b. uncomplete specification might be cumbersome to use
  b. other specifications (EMMA, SMIL) also
 deliver "complete service"

In this report we try to show both approaches for consideration.

Nearby specifications used in this report (other suggestions WELCOME):
A) HTML (Hypertext Markup Language)
html is a ML to format docs in browsers [2].
B) SMIL (Synchronized Multimedia Integration Language)
smil is a markup to format streaming media shows [3].
C) EMMA (Extensible MultiModal Annotation Markup Language)
emma is a markup describe multimodal interaction [4]. 
D) MIME types
content types defined by the MIME group include "text/html", "image/png", "image/gif", "video/mpeg", "text/css", and
"audio/basic" [6] 

Now we go through the three requirements of "links to the rest of the
1. "links to media", 2. "position on a timeline / emotion timing" and 3. "semantics of links".
Links 1. Links to media
  cite requirements [1];
  "The emotion markup must be able to refer to external media of various
  kinds, including videos, pictures, audio files, and nodes in an XML
  SUGGESTION a) recommend use of other spec:
    i) with HTML
     Link comes in two flavours, "a" and "link", the difference between <a>
     tag and <link> tag is that links may only appear in the header of a
     html doc, will not be displayed and is not allowed to have an end-tag.
     <emotion> <category name="pleasure"/>
     <html:a href="file:///mySound.pcm" type="audio/basic"/></emotion>
    ii) with SMIL
     One could use the SMIL <ref> tag and its synonyms
    "animation", "audio", "img", "text", "textstream" and "video". 
     <emotion> <category name="pleasure"/>
     <smil:audio src="file:///mySound.pcm"/> </emotion>
    iii) with EMMA
     There is no own tag for source-signals, only an attribute to the "interpretation"-tag
     <emotion> <category name="pleasure" />
      <emma:interpretation emma:signal=""
        emma:medium="acoustic" emma:mode="voice"/> </emotion>
  SUGGESTION b) define an own tag for links to media:
    Because the application of the other MLs differs from that of the EmotionML,
    we might define our own link and baptize it, perhaps, "media". 
    Of course we still can use URIs [5] and MIME types [6].
    <emotion> <category name="pleasure"/>
    <media src="file:///mySound.pcm"/ type="audio/basic"> </emotion>
Links 2. Position on a time line in externally linked objects and
  cite requirements [1]:
  "The emotion markup must provide a mechanism for complementing a link
  to media with timing information, in order to further specify the
  scope of the link."
  Together with
  Core 8. Emotion timing: "The emotion markup must provide a generic mechanism for temporal scope."
  Generally spoken we could use for durations the clock value description for
  time description, as specified in [8]. 
  An example is 
   02:30:03 = 2 hours, 30 minutes and 3 seconds or 30s = 30 seconds. 
  For absolute values we might use, like EMMA, the number of milliseconds since 1
  January 1970.

  SUGGESTION a) recommend use of other spec:
    i) with EMMA
     As Emma deals with multiple input events, its time support is
     specialized on events starting relative to other events.
     <emma:interpretation id="int1" emma:start="1087995961542" emma:end="1087995963542"/>
    ii) with SMIL
     Because SMIL is about displaying streams of multimedia objects, it
     has support for very complex timed events running in parallel or
     sequential. A very simple example is this:
     <span smil:timeAction="visibility" smil:begin="3s" smil:dur="5s">
       Show this text for a short period.
  SUGGESTION b) define own attributes for timing information
    We than have to decide whether timing shall be expressed by 
       i) an explicit own element:
           <category name="pleasure"/>
           <media src="file:///mySound.pcm" type="audio/basic"/>
           <timeSpec start="2s" duration="5s"/> 
       ii) implicitely, as a set of timing attributes that are valid for
          specific elements, like EMMA and SMIL do,
          <emotion start="2s"  end="7s" duration="5s"> 
           <category name="pleasure"/>
           <media src="file:///mySound.pcm" type="audio/basic"/>

    In order to fulfil requirement 8 iii), the referential description of
    timing (see [1]), we could use the Emma syntax: and define two
    elements: <start> and <end> that can have timing attributes.
    This would be an example for "start 2 seconds before utterance starts and
    ends with the second noun-phrase..." (with namespaces)
    <utterance id="utt1" emma:start="1231314123"/>
    <nounPhrase id="np1" emma:start="1231316123" emma:duration="3s"/>
    <emoml:emotion emma:start="2s" emma:duration="5s"> 
      <emoml:category set="everyday" name="pleasure"/>
      <emoml:start emma:time-ref-uri="#utt1"
        emma:offset-to-start="2s" />
      <emoml:end   emma:time-ref-uri="#np1"
        emma:offset-to-start="0" />
   Requirement 8 iv) 
   "a sampling mechanism providing values for variables at evenly
   spaced time intervals." 
   needs clarification, but perhaps should not be a MUST-HAVE requirement?
Links 3. The semantics of links to the "rest of the world"
  cite requirements [1];
  "The emotion markup must provide a mechanism for assigning meaning to
  those links. The following initial types of meaning are envisaged:
      * The experiencer (who "has" the emotion);
      * The observable behaviour "expressing" the emotion;
      * The trigger/cause/emotion-eliciting event of the emotion;
      * The object/target of the emotion (the thing that the emotion is about)...."
  This reminds a bit of our category discussion [9], where we also
  have the problem to allow users to define an own set but also want to
  suggest some pre-defined.
  We suggest to solve it in a similar fashion by defining an element, perhaps
  called "link", that takes as attributes
  - a role, to declare the "semantic role" of the link
  - a src, similar to audio src: the target url
  - a set, similar to the category set attribute, in order to point to
    the set of pssible semantic roles.
  Here's an example of Jane Doe who is pleased about a concert performed
  by the Beatles, described by the roles suggested in the requirements:
    <category set="everyday" name="pleasure"/>
    <link set="myEmotionRoles" role="experiencer"
    <link set="myEmotionRoles" role="behaviour"
    <link set="myEmotionRoles" role="trigger"
    <link set="myEmotionRoles" role="target"


This requirement seems to really be in 2 parts, a. the link and b. the semantics.
In this case we could mix this specification and use one of the link examples 
above but specify a new emotion tag for the semantics, which are emotion specific.

    <category name="pleasure" />
    <context role="experiencer" id="Jane Doe">
        <html:a href="" />